The goal of vsp is to enable fast, spectral estimation of latent
factors in random dot product graphs. Under mild assumptions, the vsp
estimator is consistent for (degree-corrected) stochastic blockmodels,
(degree-corrected) mixed-membership stochastic blockmodels, and
degree-corrected overlapping stochastic blockmodels.
More generally, the vsp estimator is consistent for random dot product
graphs that can be written in the form
E(A) = Z B Y^T
where Z and Y satisfy the varimax assumptions of [1]. vsp works
on directed and undirected graphs, and on weighted and unweighted
graphs. Note that vsp is a semi-parametric estimator.
You can install the development version of vsp with:
install.packages("devtools")
devtools::install_github("RoheLab/vsp")Obtaining estimates from vsp is straightforward. We recommend
representing networks as igraph objects or
(sparse) Matrix objects from the
Matrix
package (see also ?Matrix).
Once you have your network in one of these formats, you can get
estimates by calling the vsp() function. The result is an object of
class "vsp_fa".
library(Matrix)
library(vsp)
library(igraph)
data(enron, package = "igraphdata")
enron_fa <- vsp(enron, rank = 30, scale = TRUE)
enron_fa
#> Vintage Sparse PCA Factor Analysis
#>
#> Rows (n): 184
#> Cols (d): 184
#> Factors (rank): 30
#> Lambda[rank]: 0.2077
#>
#> Pre-Processing Options (TODO)
#>
#> Components
#>
#> Z: 184 x 30 [dgeMatrix]
#> B: 30 x 30 [dgeMatrix]
#> Y: 184 x 30 [dgeMatrix]
#> u: 184 x 30 [matrix]
#> d: 30 [numeric]
#> v: 184 x 30 [matrix]- Rohe, K. & Zeng, M. Vintage Factor Analysis with Varimax Performs Statistical Inference. 2020+.
Code to reproduce the results from the paper is available here.